Time for action – the second run

If we take this representation as the input to another run of the job, we would expect nodes 2, 3, and 4 to now be complete, and for their neighbors to now be in the Current state. Let's see; execute the following steps:

1. Execute the MapReduce job by executing the following command:

   $ hadoop jar graph.jarGraphPathgraphout1graphout2
   

2. Examine the output file:

   $ hadoop fs -cat /home/user/hadoop/graphout2/part-r000000
   12,3,40D
   21,41D
   31,5,61D
   41,21D
   53,62C
   63,52C
   76-1P
   

What just happened?

As expected, nodes 1 through 4 are complete, nodes 5 and 6 are in progress, and node 7 is still pending, as seen in the following figure:

If we run the job again, we should expect nodes 5 and 6 to be Done and any unprocessed neighbors to become Current.